tests/testthat/test-xmr_recalc_shortruns_4.R

In xmrr: Generate XMR Control Chart Data from Time-Series Data

context("Calculate XMR Short Run Recalculation for 3/4 points")

library(testthat)
library(dplyr)
library(tidyr)

Measure <- c(58, 57, 69, 62, 66, 58, 66, 52, 67, 51, 51)
Time <- c(2000:2010) 
example_data <- data.frame(Time, Measure)

df <- xmr(example_data, measure = "Measure", recalc = T)

test_that("Lower shortrun recalculation is correct", {
  mv <- df$`Moving Range`[df$Order %in% c(8, 10, 11)] %>% mean()
  avm <-  df$`Average Moving Range`[8]
  calc <- avm - mv
  max <- max(calc, na.rm = T)
  expect_lt(max, 0.01)
  
  m <- df$`Measure`[df$Order %in% c(8, 10, 11)] %>% mean()
  cl <-  df$`Central Line`[8]
  calc <- m-cl
  max <- max(calc, na.rm = T)
  expect_lt(max, 0.01)
})

test_that("Upper Process Limit calculation is correct", {
  up <- df$`Upper Natural Process Limit` 
  calc <- (df$`Central Line` + (2.66 * df$`Average Moving Range`))
  max <- max(up - calc, na.rm = T)
  expect_lt(max, 0.01)
})

test_that("Lower Process Limit calculation is correct", {
  lower <- df$`Lower Natural Process Limit` 
  calc <- (df$`Central Line` - (2.66 * df$`Average Moving Range`))
  calc <- ifelse(calc < 0, 0, calc)
  max <- max(lower - calc, na.rm = T)
  expect_lt(max, 0.01)
})


Measure <- c(58, 57, 69, 62, 66, 58, 76, 73, 55, 72)
Time <- c(2000:2009) 
example_data <- data.frame(Time, Measure)

df <- xmr(example_data, measure = "Measure", recalc = T, shortrun = c(3,4))

xmr_chart(df, "Time", "Measure") + 
  geom_text(aes(y = Measure, label = Order), vjust = -1)

test_that("Upper shortrun recalculation is correct", {
  mv <- df$`Moving Range`[df$Order %in% c(7, 8, 10)] %>% mean()
  avm <-  df$`Average Moving Range`[7]
  calc <- avm - mv
  max <- max(calc, na.rm = T)
  expect_lt(max, 0.01)
  
  m <- df$`Measure`[df$Order %in% c(7, 8, 10)] %>% mean()
  cl <-  df$`Central Line`[7]
  calc <- m-cl
  max <- max(calc, na.rm = T)
  expect_lt(max, 0.01)
})


test_that("Upper Process Limit calculation is correct", {
  up <- df$`Upper Natural Process Limit` 
  calc <- (df$`Central Line` + (2.66 * df$`Average Moving Range`))
  max <- max(up - calc, na.rm = T)
  expect_lt(max, 0.01)
})

test_that("Lower Process Limit calculation is correct", {
  lower <- df$`Lower Natural Process Limit` 
  calc <- (df$`Central Line` - (2.66 * df$`Average Moving Range`))
  calc <- ifelse(calc < 0, 0, calc)
  max <- max(lower - calc, na.rm = T)
  expect_lt(max, 0.01)
})